ROS Integration of External Vehicle Motion Simulations with an AIMSUN Traffic Simulator as a Tool to Assess CAV Impacts on Traffic

A challenge with predicting the in-traffic performance of Connected and Autonomous Vehicles (CAV) is that CAV algorithms are often analyzed on a per-vehicle basis, but their effects and interactions with surrounding traffic require analysis of traffic-network behaviors. The tools for CAV simulations generally encompass two domains: 1) traffic micro-or macro-simulations which encompass traffic laws and large groups of vehicles guided by simple behavioral algorithms, and 2) on-vehicle system simulations enabling the complex algorithms for sensing and control within the immediate vicinity of the ego-vehicle. In this paper, an example is presented that bridges these two tools. Specifically, AIMSUN, a traffic modeling and traffic network simulation tool, is integrated with Robot Operating System (ROS), an open-source meta-operating system, to develop a co-simulation platform bridging traffic simulations with ego-vehicle CAV simulations. Establishing such a co-simulation platform requires a bi-directional data-flow bridge between the two software platforms wherein the motion of the ego-vehicle at each time-step in ROS is a function of the traffic scenarios as simulated by AIMSUN. User Datagram Protocol (UDP) which allows for large amounts of data transmission with low latency is used as the communication protocol for the bridge. The time latency of the bridge is analyzed by performing a loop-back test and obtaining the time delay statistics. A step-by-step tutorial is presented in this paper to guide the reader through the process of implementing such a bridge within a driving simulator environment. The co-simulation platform is demonstrated through an application example where a user can virtually drive an ego vehicle through an AIMSUN traffic network, and the co-simulation behavior is assessed by the Time-To-Collision (TTC) parameter.